It's true we need different criteria in different circumstances. For A/B tests where the goal is to help users ask questions that the community is interested in answering, the quality of answers is much less important than the mere presence of answers. And if we have the stronger signal of voting, we listen to that first. I think your point about downvoted and answered questions being interesting is on-point. Arguably questions with some downvotes are more interesting than those without since controversial posts tend to gather downvotes. (Glad you enjoyed the discussion!)

Answers can also indicate that the question was trivial to answer, making it an attractive easy rep source for users trying to establish themselves. On SO, these are often poor questions lacking even basic research, even often being duplicates.

@MonicaCellio No, the answers are often fairly long. On SO, they typically require a lengthy code block. Come to think of it, I often see high reputation users answering them as well.

@jpmc26: It sounds like you are bothered less by the answers themselves and more by illicite reputation gain and incentivising otherwise poor questions.

@JonEricson It's really about turning SO into a personal help desk rather than a quality documentation resource that bothers me. Answering questions where the author doesn't have a good grasp of their problem or hasn't even tried to struggle through finding an answer goes a very long way toward doing that. The point in the context here about measuring question quality is that having answers indicates nothing about how much homework the author did beforehand. In fact, I wouldn't be surprised if the opposite was true, that questions where the author has done that hard work have fewer answers.

@jpmc26: Presumably authors who do ask difficult questions will at least get upvotes, which is a stronger signal than answers. And questions that get answered and downvoted are also counted as bad under the question grade system. It's only when we don't have voting that an answer matters at all. And considering the drop in answering on Stack Overflow, that's less and less likely to be relevant. :-/

@jpmc26: That's an amazing question with a wonderful answer. They both got my vote. Thanks for pointing it out. Makes me want to go back to using Python instead of Ruby. Now what's the problem with it?

@JonEricson Read the comments. It's actually unclear what they're asking because they didn't specify a runtime and apparently didn't check the docs to see whether the behavior is guaranteed or not. (You know this because of their response that they were asking about the language rather than the runtime, meaning they didn't verify whether it's actually required or not.) It's also something that doesn't even really have an answer, other than "some guy apparently chose it at random." It's also useless click bait, since hash values are, by definition, arbitrary and subject to change.

@JonEricson To put it another way, it's an excellent example of a common attitude: "I want some information and I'm not going to bother looking into it before I ask. Not even to fact check my thoughts about it, make sure I have some grasp of the situation, or try to find an answer in the obvious place." There's no excuse for a 172k rep user with a gold badge in Python to not delve into those issues on their own.

@JonEricson The site is flooded with questions asked from that mindset, and that's the difference between being a help desk where you're just doing other people's work for them and a documentation site where you're putting information into a form that's easier for people to find and understand. No one wants to spend the time to learn about their problem or its possible solutions. They just want to be handed the final answer on a silver platter.

@jpmc26: The asker wants the final answer on a silver platter because they didn't find a nineteen year old commit? When you asked which runtime and the author had already tested three. Your comment says the value is not documented but you complain people don't read the documentation. If the site is flooded with such questions, why did you pick such a poor example to demonstrate your point? This is about as opposite of treating the site as a helpdesk as I can imagine. If you still have a problem with it, I suggest a discussion question on MSO.

@JonEricson No, they want the answer on a silver platter because they didn't check and document two obvious things that are well within their experience level to provide context: 1. Is the behavior the same on all implementations? If not, which one(s) are you asking about? 2. What does the documentation say about the behavior? I made this perfectly clear; you can see it in my previous comments. Do not misrepresent my position, as SO staff commonly does. Notably, the commit you mention doesn't even provide an actual answer as to why an approximation of 10000*pi was chosen, anyway.

@JonEricson "Your comment says the value is not documented but you complain people don't read the documentation." No. I'm complaining that the asker didn't check whether the value and the reason for it were documented. I know they didn't because they specifically said (in a now deleted comment) that they were asking about the language specification. The spec doesn't define the value, meaning that it's not valid to ask about the value as part of the language spec. wim should have discovered that by checking the documentation beforehand and narrowed the question to specific runtime(s).

@JonEricson I suspect the difference may still be difficult for you to wrap your head around, so let me put it one more way: if they had checked the documentation at all, they would have known the value was not part of the language specification. Therefore, since they did not know it wasn't part of the language spec, they must not have checked the documentation at all. Findings like that also need to be documented in the question itself, to clarify the nature of a question. Asking about the value as part of the language specification shows wim did not have a good grasp of the subject.

@JonEricson And now that you completely misrepresented what I was saying, several other people have readily latched onto that misinterpretation as valid criticism. There is nothing more unwelcoming than knowing no one is going to actually listen to what you're saying carefully enough to try to understand it.

Especially when you've put as much effort as you can into being as crystal clear as you can because you know the other person isn't going to readily agree.

Careful thought and effort aren't rewarded here. That's the fundamental problem. I know I'm not always right, but I'm not always wrong, either. I work very hard to try to understand whatever subject I'm examining, whether it's a Meta issue or a main topic, and I work equally hard to articulate my thoughts.

But it doesn't make a difference, because very few people are even interested in trying to develop their understanding or engage with someone else's. So it's mostly a waste. It's easier to malign someone who disagrees with you than engage in an actual discussion.